Genes for napping, not naps themselves, were linked to dementia in Swedish twins

Poor sleep and dementia often travel together, but untangling which comes first is hard. Sleep changes can be an early sign of a brain disease that has already started, and people who sleep badly often differ from good sleepers in many other ways. A team at Karolinska Institutet in Sweden tried two tricks at once to get around this: comparing twins with each other, and looking at the genes that nudge people towards certain sleep habits.

The result is a split picture. Napping and longer sleep in old age were linked to lower dementia risk, yet a genetic tendency to nap was linked to higher risk.

What they did

The researchers used the Swedish Twin Registry, which covers nearly all twins born in Sweden. Between 1998 and 2002, twins born in 1958 or earlier were interviewed by phone; some also answered a validated sleep questionnaire about how long they slept, insomnia, snoring, daytime sleepiness, whether they were morning or evening types, and whether they napped. Dementia diagnoses were taken from national registers of hospital visits, death records and prescriptions, up to the end of 2016.

The analysis had three parts:

• Reported sleep. 13,087 twins with questionnaire answers, average age 70, followed for about 13 years; 1,541 developed dementia.
• Genetic sleep scores. For 16,164 twins who had given blood or saliva, the team calculated a polygenic score for each sleep habit: a sum of many small genetic variants that, in a large British study, went with that habit. These people were younger (average age 55) and followed for about 16 years; 559 developed dementia.
• Both. 3,133 twins had both kinds of data; 377 developed dementia.

To reduce the chance that early, undiagnosed disease was changing sleep, the main analyses left out anyone diagnosed in the first 10 years of follow-up. The team also compared twins within the same pair. Twins share family background and, in part or in full, their genes, so if a sleep habit still goes with dementia when one twin is compared with the other, family and genetic factors are less likely to explain it.

What they found

• Longer sleep was linked to slightly lower risk (hazard ratio 0.93, with a confidence interval reaching 1.00, so only just at the edge of chance). Napping was also linked to lower risk at first, but this faded after adjusting for behaviour and health.
• Within twin pairs, both links disappeared (hazard ratios 1.03 for sleep length and 1.02 for napping). The researchers read this as a sign that shared genes or family background, rather than the sleep habit itself, were behind the links.
• The genetic score for napping pointed the other way. Among genotyped twins, each standard-deviation step up in the napping score was linked to a 12% higher dementia risk (hazard ratio 1.12, interval 1.01 to 1.23). It was the only sleep score with a clear link.
• This held when actual napping was taken into account. In the 3,133 people with both kinds of data, the napping score was still linked to higher risk (hazard ratio 1.17) after adjusting for whether they really napped, and the two did not interact.
• Accounting for deaths before any dementia diagnosis changed the picture little.

What to keep in mind

The authors list several limits. All sleep data were self-reported, which invites recall errors. Napping was only a yes-or-no question, so length, timing and frequency of naps could not be studied, and no genetic study has yet found variants for those details. The team could not use the largest available genetic study of sleep, which may weaken some scores, and some questions were worded differently from the British study the scores came from (for example "heavy snoring" versus "snoring"). There were too few twin pairs to compare identical and non-identical twins separately. Registers can also miss early or mild dementia.

The genetic scores also have a catch the authors point out themselves: many napping-related variants are also linked to blood pressure, body weight and type 2 diabetes, which are known dementia risk factors. So the napping score may be a marker of broader biological vulnerability rather than a sign that napping itself does harm. The differences in age matter too: the genetic scores reflect midlife sleep habits, while the questionnaire captured sleep at around 70. A possible interaction between daytime sleepiness and its genetic score was seen, but the authors ask for caution because it rests on small numbers and wide intervals.

Why it matters

Studies of sleep and dementia keep producing mixed results, and this one shows why. A habit and the genes behind it can point in opposite directions. The researchers suggest that napping in late life may be a normal or even restorative part of aging, whereas an inherited tendency to nap may flag a separate vulnerability that shows up decades later.

The twin comparison adds a sobering note for headlines about sleep length: the link with longer sleep did not survive comparing siblings who share a family. The authors call for genetic tools that capture when and how long people nap, and for repeating the analysis in larger groups.